W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > July 2003

Re: Does XSD/RNG Based RDF in XHTML need a new FPI?

From: Joseph Reagle <reagle@w3.org>
Date: Fri, 4 Jul 2003 09:09:40 -0400
To: Masayasu Ishikawa <mimasa@w3.org>, public-rdf-in-xhtml-tf@w3.org
Cc: Dan Connolly <connolly@w3.org>
Message-Id: <200307040909.40560.reagle@w3.org>

On Friday 04 July 2003 06:51, Masayasu Ishikawa wrote:
> XHTML+CC, but defining XHTML+RDF is like defining XHTML+arbitraryXML.

Yes, exactly. So what is necessary in this case? You had mentioned this 
might be considered a "not-so-strictly" conforming XHTML document.

> If you stick to the DTD, it's just invalid, regardless of the XHTML
> host language document type conformance.

Ok, so I remove the DTD.

> You cannot claim it's
> XHTML 1.0 Transitional document type (which is a specific XHTML Family
> member with fixed set of elements and attributes).

I've been trying to follow this [1] thread, and [2], but I'm still not sure 
if this concept of "not-so-strictly" is a red-herring I'm chasing, nor how 
to properly identify a "not-so-strictly" document of  XHTML+arbtraryXML. In 
the past, my understanding was that an XML instance was self identifying 
via its DOCTYPE and/or namespace. DOCTYPEs are related to DTDs, so let's 
put that aside. Then the namespace should be able to identify the 
instance's definition. However, this itself is a contentious statement as 
some folks feel that a namespace merely disambiguates, and is not 
self-defining nor de-referencable. Ok, so what do we rely upon so that a 
XML document can properly identify itself for validation? That seems to be 
specific to what you mean by validation. In this Schema case, there's the 
schemaLocation stuff. However, these attributes are not permitted in XHTML 
and you noted that's not likely [4] to happen any time soon?

Now, there's NRL. I've never had an expectation that one could or should 
address the semantics of composing/mixing XML1.0 . XML 1.0 just wasn't made 
for that, and I don't think it will be able to be added layer. (So I think 
this issue will be on the TAG list for quite some time [6] <smile/>). But I 
did hope that one could have the syntax be self-identifying and 
validatable. NRL seems to put an end to that argument. The namespaces 
disambiguate the elements and attributes, that's all. The conception of 
validity is an external definition without any way to necessarily identify 
that definition from an instance. (This, and playing with various schema 
tools this week for something else leads me to feel we're headed towards 
"catalogs" in the XML space.)

Ok, so sorry for the rambling, but to put a point on my question, could you 
point me to a proper instance that you think conforms to your [3]. Does [a] 
without the DOCTYPE qualify?

Received on Friday, 4 July 2003 09:09:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:50:17 UTC